基于信息熵的容器云平台监控模型设计

Design of Monitoring Model of Container Cloud Platform Based on Information Entropy

为了提高容器云平台监控和数据调度分配能力,提出基于信息熵的容器云平台监控模型。采用资源冗余度分析和CPU物理核心资源检测方法,提取容器云平台监控的信息熵,采用负载均衡控制的方法,分析容器云平台的负载变化和资源调度延迟,将容器云平台的运行数据分为决策类数据、计算资源池数据以及监控服务数据类,采用类型化的特征匹配和数据聚类方法,建立器云平台监控模型的信息熵和关联信息特征匹配模型,通过多个复杂均衡的任务调度序列进行随机链路动态分配,采用容器云的任务排队模型,实现容器云平台监控模型的优化设计。仿真测试结果表明,采用该方法进行容器云平台监控的云数据调度能力较好,资源利用率达到100%,收敛值达到12 000以上,提高了资源利用率和云资源的在线调度能力。

In order to improve the monitoring and data scheduling ability of container cloud platform, a monitoring model of container cloud platform based on information entropy is proposed. Using resource redundancy analysis and CPU physical core resource detection method, the information entropy of container cloud platform monitoring is extracted, the load change and resource scheduling delay of container cloud platform are analyzed by load balancing control method, and the operation data of container cloud platform is divided into decision-making data, computing resource pool data and monitoring service data. The information entropy and associated information feature matching model of container cloud platform monitoring model is established by using typed feature matching and data clustering method, and the container cloud platform is realized by dynamically allocating multiple complex and balanced task scheduling sequences and adopting the task queuing model of container cloud. The simulation test results show that the cloud data scheduling ability of container cloud platform monitoring using this method is good, the resource utilization rate reaches 100%, and the convergence value reaches more than 12 000, which improves the resource utilization rate and the online scheduling ability of cloud resources.